CN105203543B - FCL red grape fruit size fractionation devices and methods therefor based on machine vision - Google Patents
FCL red grape fruit size fractionation devices and methods therefor based on machine vision Download PDFInfo
- Publication number
- CN105203543B CN105203543B CN201510607956.2A CN201510607956A CN105203543B CN 105203543 B CN105203543 B CN 105203543B CN 201510607956 A CN201510607956 A CN 201510607956A CN 105203543 B CN105203543 B CN 105203543B
- Authority
- CN
- China
- Prior art keywords
- image
- red grape
- fruit
- fcl
- grape fruit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Sorting Of Articles (AREA)
Abstract
The invention discloses a kind of FCL red grape fruit size self-grading device and its method based on machine vision, it is related to colony's fruit grading technical field.This method is：1. FCL red grape to be measured is numbered and placed on conveyer belt；2. FCL red grape being sent into seal box carry out IMAQ successively, while obtain near-infrared image and coloured image；3. image procossing is carried out to near-infrared image and coloured image：Read image；Image is pre-processed；Image co-registration；Edge extracting；Grape fruit image is split；Remove small area processing；Ellipse fitting；Extracting parameter；Ellipse is screened；Establish hierarchy model and examine.The present invention utilizes machine vision technique, can complete the automatic classification of FCL red grape；Grade scale is unified, and classification accuracy is high；Non-Destructive Testing is realized using image processing techniques, and detection speed is fast；Size Non-Destructive Testing is realized to group object with machine vision technique, had broad application prospects.
Description
Technical field
The present invention relates to colony's fruit grading technical field, more particularly to a kind of FCL red grape based on machine vision
Fruit size self-grading device and its method.The present invention mainly using industrial camera collection red grape image, passes through one
Series image processing method, extract red grape fruit size parameter, according to the size of fruit establish model and examine into
Row classification, a kind of lossless quick method is provided for the automatic classification of red grape.
Background technology
At present, the classification of red grape fruit size passes through several of shear surface mainly by manually come what is realized
Fruit, by the use of its fruit maximum axial diameter of vernier caliper measurement as the diameter of fruit, according to the diameter for several fruits surveyed
To judge the red grade for putting forward fruit of FCL.During for red grape fruit size fractionation, it is necessary to follow 15% degradation principle；Change
Yan Zhi, if the two level fruit for being detected red grape fruit is on the high side, but when the three-level fruit accounting being detected in red grape fruit surpasses
Cross 15%, then be detected red grape fruit and be reduced to three-level fruit.During manual grading skill, Finite Samples, maximum axial diameter are chosen
Bad assurance, manual grading skill intricate operation, efficiency are low and with destructiveness in addition.
Machine vision is widely used in grading field, such as the application of apple, pears, egg etc..Studies in China
Research object is mainly used as using the regular shapes such as apple, citrus, egg and segregative single body.And for similar red
Grape is this to be not readily separated, shape size differs, the research of compact-sized object is very few.
It is numerous come the research of detection object size with machine vision.Xiong Lirong etc. to the pixel of simple grain peanut by carrying out
Extraction and calculating, realize the classification of simple grain peanut size【Xiong Lirong, a chess-playing circles, Xiao Renqin, the peanut based on machine vision are big
Small inspection, hubei agricultural science [J], 2007,46 (3)】；Chen Hong etc. proposes a kind of sorting side of curvilinear structures signature analysis
Method, devise a kind of flower mushroom separation system based on machine vision【Chen Hong, Xia Qing, left booth etc., the flower mushroom based on machine vision point
Selecting technology, agricultural mechanical journal [J], 2014,45（1）】；Deng Haixia is in colony's birds, beasts and eggs size detection method based on machine vision
The barycenter of birds, beasts and eggs is first tried to achieve in research, then seeks major axis of the barycenter to the maximum of marginal point as birds, beasts and eggs, minimum value is as birds, beasts and eggs
Short axle【Deng Haixia, birds, beasts and eggs size detection method research [D] Hua Zhong Agriculture University of colony based on machine vision, 2006】.Though
The research that right machine vision detects to article size is numerous, but primarily directed to easy segmentation or the thing occurred with single body
Body carries out detection classification, is studied for objects such as the red grapes into string-like few.
Red grape using machine vision to red grape fruit size research primarily directed to whole string.For example, Chen Ying,
Liao Tao et al. calculates the size and shape parameter of the whole string of grape using perspective plane area method and fruit direction of principal axis drop shadow curve, and then realizes
To the quality grading of grape, 90% is up to the accuracy rate of color grading, the Grading accuracy rate of size shape is then reached
88.3%【Leaned at the beginning of Chen Ying, Liao Tao, woods etc., the grape detection hierarchy system based on machine vision, agricultural mechanical journal [J] .2010,
41（3）】.Wang Qiaohua etc. has invented a kind of red grape self-grading device based on machine vision and whole string red grape has been carried out
Detection, by extracting the size characteristic of red grape fruit the methods of the angle of curvature and catastrophe point that calculate red grape, classification is just
True rate is 85%【Wang Qiaohua, Ding Youchun, Luo Jun, permitted that Kun is auspicious, Li Min, red grape self-grading device based on machine vision and
Its method [P], Chinese patent：CN102680414A, 2012-09-19】.
It is domestic at present it is not yet found that closing FCL red grape fruit size automatic classification technology and screening installation through retrieval
Applied to produce reality.
The content of the invention
The purpose of the present invention is that the problem of overcoming prior art to exist, there is provided a kind of FCL based on machine vision is red
Grape fruit size self-grading device and its method.
The object of the present invention is achieved like this：
The present invention utilizes machine vision, and by specific Processing Algorithm, FCL red grape fruit is split, then profit
The size parameter of red grape fruit is extracted with ellipse fitting scheduling algorithm, and then realizes the size point of FCL red grape fruit
Level；The present invention entirely automates into fruit classification process again from IMAQ to image procossing, and grade scale is unified, point
Step velocity is fast, and accuracy rate is high, has important application meaning.
Present invention is generally directed to FCL red grape fruit size to be analyzed, unlike other existing researchs,
The present invention is studied for the red grape of FCL, and not directed to single red grape of going here and there.In the present invention, main completion is following
Several work：
1st, image collecting device is built, FCL red grape image is gathered, mainly includes the selection of red grape,
The selection of camera, the selection of light source, collection image etc.；
2nd, using suitable image processing method, the image of collection is handled, extracts red grape fruit
Characteristic parameter：
This part is the most key link, mainly there is three steps, and first is image preprocessing, selects suitable algorithm
Remove red grape carpopodium；Second is segmentation grape fruit, and the present invention is split red grape fruit using image segmentation；
3rd is parameter extraction and conversion, and the present invention is fitted using ellipse fitting to red grape fruit, extracts red grape fruit
The parameter of grain size, while by suitable conversion algorithm, obtain the diameter of red grape fruit.
3rd, red grape fruit size automatic classification model is established, and is examined.
How emphasis and the difficult point of the present invention be in removing carpopodium and how to obtain from FCL red grape image red
The parameter of grape fruit size.The purpose to be realized is to extract actual parameter by machine vision and image processing techniques, is built
Vertical FCL red grape size fractionation hierarchy model, realizes the Fast nondestructive evaluation to FCL red grape and classification.
Specifically, technical scheme is as follows：
First, the FCL red grape fruit size self-grading device based on machine vision（Abbreviation device）
The target of the present apparatus is red grape；
It is provided with display screen, main frame, industrial camera, annular light source, seal box, fruit case and conveyer belt；
Its position and annexation are：
Passage and installation conveyer belt are opened up in seal box bottom, puts fruit case on a moving belt, red grape is positioned over
In fruit case；
A circular hole is opened up among the top of seal box, annular light source is built-in with circular, is installed at circular hole
There are industrial camera, industrial camera alignment red grape collection information；
Industrial camera is connected with main frame, real-time update camera information.
2nd, the FCL red grape fruit size automatic grading method based on machine vision（Abbreviation method）
This method comprises the following steps：
1. FCL red grape to be measured is numbered and placed on conveyer belt；
2. FCL red grape being sent into seal box carry out IMAQ successively, while obtain coloured image and near-infrared
Image；
3. image procossing is carried out to coloured image and near-infrared image；
A, coloured image is switched into HSV images（H represents tone, and S represents saturation degree, and V represents brightness）, select suitable threshold
Value is handled, and obtains removing the image after carpopodium, then be translated into binary image；
B, enhancing processing is carried out to near-infrared image, obtains clear-cut gray level image；
C, near-infrared image after enhancing is multiplied with removing the binary image of carpopodium, obtains removing the gray-scale map after carpopodium
Picture；
D, rim detection is carried out to the gray level image after removal carpopodium；
E, the edge of red grape fruit is split；
F, small area processing is removed to the edge of red grape fruit；
G, ellipse fitting is carried out to the edge of the red grape fruit after removing small area processing using ellipse fitting；
H, extracting parameter, red grape fruit is classified.
The present invention operation principle be：
FCL red grape is placed on a moving belt, seal box bottom is sent to by conveyer belt, switch on power switch,
Annular light source is opened, image capture software is opened in display screen, adjusts industrial camera parameter, control industrial camera collection FCL
The image of red grape；Image processing software passes through above method start to process FCL red grape image, Quick FCL
The grade of red grape, so as to realize the automatic detection of FCL red grape fruit size.
The present invention has the advantages that：
1. utilizing machine vision technique, the automatic classification of FCL red grape can be completed；
2. grade scale is unified, classification accuracy is high；
3. Non-Destructive Testing is realized using image processing techniques, and detection speed is fast；
4. realizing size Non-Destructive Testing to group object with machine vision technique, have broad application prospects.
Brief description of the drawings
Fig. 1 is the structural representation of the present apparatus,
In figure：
0-red grape；
1-display screen；
2-main frame；
3-industrial camera；
4-annular light source；
5-seal box；
6-fruit case；
7-conveyer belt
Fig. 2 is the workflow diagram of image processing software.
Embodiment
Describe in detail with reference to the accompanying drawings and examples：
First, device
1st, it is overall
Such as Fig. 1, the target of the present apparatus is red grape 0；
It is provided with display screen 1, main frame 2, industrial camera 3, annular light source 4, seal box 5, fruit case 6 and conveyer belt 7；
Its position and annexation are：
Passage and installation conveyer belt 7 are opened up in the bottom of seal box 5, puts fruit case 6 on conveyor belt 7, red grape 0 is put
It is placed in fruit case 6；
A circular hole is opened up among the top of seal box 5, annular light source 4 is built-in with circular, is pacified at circular hole
Equipped with industrial camera 3, industrial camera 3 is directed at red grape 0 and gathers information；
Industrial camera 3 is connected with main frame 2, real-time update camera information.
2nd, functional block
1）Display screen 1
Display screen is a kind of general outsourcing piece, such as selects the LED display on Great Wall.
2）Main frame 2
Main frame 2 is a kind of general outsourcing piece, such as selects CPU Intel Core i5-3210M 2.50GHz/ internal memory 8G,
7 64 systems of Windows；
It is embedded with capture card and image processing software.
Such as Fig. 2, the workflow of described image processing software is：
1. read image -201；
2. image is carried out to pre-process -202
By carrying out HSVization, Threshold segmentation and binary conversion treatment to coloured image, obtain removing the binary picture of carpopodium
Picture；Near-infrared image is handled simultaneously, enhancing processing is carried out to near-infrared image, obtains the near-infrared image that profile becomes apparent from；
3. image co-registration -203
The binary image for removing carpopodium is blended with enhanced near-infrared image, obtains removing carpopodium and edge obtains
To the gray level image of enhancing；
4. edge extracting -204
Rim detection is carried out to the image of image co-registration using Canny operators, obtains red grape edge；
5. grape fruit image segmentation -205
Watershed segmentation processing is carried out to gray level image, the center of red grape fruit is extracted, according to red grape
The center of fruit is split to red grape edge, obtains individual particle red grape fruit edge；
6. remove small area processing -206
Small area processing is removed to the edge of red grape fruit；
7. ellipse fitting -207
Ellipse fitting -207 is carried out to the edge for removing the red grape fruit after small area is handled；
8. extracting parameter -208
Extract oval major axis Ra, short axle Rb and eccentricity e=Ra/Rb；
9. ellipse is carried out to screen -209
As eccentricity e >=0.75, retain the ellipse, work as e<0.75, cast out the ellipse；
10. establish hierarchy model and examine -210
All red grape fruit sizes are calculated, hierarchy model is established and examines.
3）Industrial camera 3
Industrial camera 3 is a kind of general outsourcing piece, such as selects the 2-CCD cameras of JAI companies；
Its function is to gather the view data of red grape 0.
4）Annular light source 4
Annular light source 4 is made up of annular daylight lamp and near-infrared light source, is general outsourcing piece, is such as given birth to from colon company
The near-infrared light source of annular daylight lamp and the OPT production of production；
Its function is to provide light source for red grape 0.
5）Seal box 5
Seal box size is：Length × width × height=500mm × 500mm × 1000mm.
6）Fruit case 6
Fruit case 6 is a kind of Universal box, such as grape fruit case, for containing red grape.
7）Conveyer belt
Conveyer belt is a kind of general part, installs suitable conveyer belt according to actual conditions, its effect is for transmitting fruit
Case.
3rd, testing result
The test specimen of this experiment is Xinjiang red grape, and 38 case, purchases hundred supermarket in Hua Zhong Agriculture University altogether,
38 casees red grapes are tested by above-mentioned FCL red grape fruit size fractionation devices and methods therefor, wherein correctly
It is classified 35 casees, Grading accuracy rate 92.1%.
Claims (1)
1. a kind of FCL red grape fruit size automatic grading method based on machine vision, it is red that device, which includes target,
Grape (0)；
Display screen (1), main frame (2), industrial camera (3), annular light source (4), seal box (5), fruit case (6) and conveyer belt are set
(7)；
Its position and annexation are：
Passage and installation conveyer belt (7) are opened up in seal box (5) bottom, fruit case (6), Hong Ti Portugals are put on conveyer belt (7)
Grape (0) are positioned in fruit case (6)；
A circular hole is opened up among the top of seal box (5), annular light source (4) is built-in with circular, is pacified at circular hole
Equipped with industrial camera (3), industrial camera (3) alignment red grape 0 gathers information；
Industrial camera (3) is connected with main frame (2), real-time update camera information；
Described main frame (2) is embedded with capture card and image processing software, and its workflow is：
1. read image (201)；
2. (202) are pre-processed to image
By carrying out HSVization, Threshold segmentation and binary conversion treatment to coloured image, obtain removing the binary image of carpopodium；Together
When handle near-infrared image, enhancing processing is carried out to near-infrared image, obtains the near-infrared image that profile becomes apparent from；
3. image co-registration (203)
The binary image for removing carpopodium is blended with enhanced near-infrared image, obtains removing carpopodium and edge is increased
Strong gray level image；
4. edge extracting (204)
Rim detection is carried out to the image of image co-registration using Canny operators, obtains red grape edge；
5. grape fruit image splits (205)
Watershed segmentation processing is carried out to gray level image, the center of red grape fruit is extracted, according to red grape fruit
Center red grape edge is split, obtain individual particle red grape fruit edge；
6. remove small area processing (206)
Small area processing is removed to the edge of red grape fruit；
7. ellipse fitting (207)
Ellipse fitting is carried out to the edge for removing the red grape fruit after small area is handled；
8. extracting parameter (208)
Extract oval major axis Ra, short axle Rb and eccentricity e=Ra/Rb；
9. (209) are screened to ellipse
As eccentricity e >=0.75, retain the ellipse, work as e<0.75, cast out the ellipse；
10. establish hierarchy model and examine (210)
All red grape fruit sizes are calculated, hierarchy model is established and examines；
It is characterized in that comprise the following steps：
1. FCL red grape to be measured is numbered and placed on conveyer belt；
2. FCL red grape being sent into seal box carry out IMAQ successively, while obtain coloured image and near-infrared figure
Picture；
3. image procossing is carried out to coloured image and near-infrared image；
A, coloured image is switched into HSV images, selects suitable threshold value to be handled, obtain removing the image after carpopodium, then will
It is converted into binary image；
B, enhancing processing is carried out to near-infrared image, obtains clear-cut gray level image；
C, near-infrared image after enhancing is multiplied with removing the binary image of carpopodium, obtains removing the gray level image after carpopodium；
D, rim detection is carried out to the gray level image after removal carpopodium；
E, the edge of red grape fruit is split；
F, small area processing is removed to the edge of red grape fruit；
G, ellipse fitting is carried out to the edge of the red grape fruit after removing small area processing using ellipse fitting；
H, extracting parameter, red grape fruit is classified.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510607956.2A CN105203543B (en) | 2015-09-22 | 2015-09-22 | FCL red grape fruit size fractionation devices and methods therefor based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510607956.2A CN105203543B (en) | 2015-09-22 | 2015-09-22 | FCL red grape fruit size fractionation devices and methods therefor based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105203543A CN105203543A (en) | 2015-12-30 |
| CN105203543B true CN105203543B (en) | 2018-04-10 |
Family
ID=54951337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510607956.2A Active CN105203543B (en) | 2015-09-22 | 2015-09-22 | FCL red grape fruit size fractionation devices and methods therefor based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105203543B (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105606473B (en) * | 2016-01-20 | 2018-06-15 | 华中农业大学 | A kind of red grape rigidity nondestructive testing method based on machine vision |
| CN105699384A (en) * | 2016-02-01 | 2016-06-22 | 柳州市诚明科技有限公司 | Silkworm cocoon quality detection device |
| CN105698839A (en) * | 2016-02-01 | 2016-06-22 | 柳州市诚明科技有限公司 | Method for detecting silkworm cocoon quality |
| CN105738294B (en) * | 2016-03-01 | 2018-11-06 | 江苏大学 | The spike fruit automatic detection device and method of imaging are regarded based on monocular more |
| CN106111556B (en) * | 2016-08-22 | 2018-02-23 | 华中农业大学 | The automatic detection of spike fruit is equipped with classification |
| CN106666767B (en) * | 2016-11-18 | 2019-04-12 | 大连民族大学 | A kind of efficient sunflower seeds hulling method of view-based access control model technology |
| CN108318494B (en) * | 2018-05-15 | 2019-10-08 | 华中农业大学 | The red online vision detection and classification devices and methods therefor for proposing fruit powder |
| CN109724978B (en) * | 2019-01-02 | 2021-06-01 | 广州南颐贸易有限公司 | Accurate intelligent classifying and selecting device for swallow cup |
| CN109856144A (en) * | 2019-03-13 | 2019-06-07 | 西北农林科技大学 | A kind of Kiwi berry based on mobile phone expands fruit detection method and device |
| CN110064601B (en) * | 2019-05-23 | 2023-07-18 | 仲恺农业工程学院 | Seedling detection and classification system and classification method for vegetable grafting |
| CN110174065B (en) * | 2019-06-17 | 2021-02-26 | 湖南农业大学 | Fruit size nondestructive testing method based on orthogonal binocular machine vision |
| CN110567968A (en) * | 2019-09-04 | 2019-12-13 | 中国科学院长春光学精密机械与物理研究所 | part defect detection method and device |
| CN114179056B (en) * | 2021-12-13 | 2023-07-25 | 国网湖南省电力有限公司 | GIS indoor multi-degree-of-freedom rail-mounted inspection robot and application method thereof |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102680414B (en) * | 2012-05-30 | 2014-06-25 | 华中农业大学 | Automatic grading device for red globe grapes based on machine vision and method thereof |
| CN103792235A (en) * | 2014-01-10 | 2014-05-14 | 内蒙古农业大学 | Diffuse transmission spectrum and image information fusion method for detecting internal quality of honeydew melons on line and device |
-
2015
- 2015-09-22 CN CN201510607956.2A patent/CN105203543B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105203543A (en) | 2015-12-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105203543B (en) | FCL red grape fruit size fractionation devices and methods therefor based on machine vision | |
| Sahu et al. | Defect identification and maturity detection of mango fruits using image analysis | |
| Liming et al. | Automated strawberry grading system based on image processing | |
| Blasco et al. | Automatic sorting of satsuma (Citrus unshiu) segments using computer vision and morphological features | |
| US9527115B2 (en) | Computer vision and machine learning software for grading and sorting plants | |
| CN110276386A (en) | A kind of apple grading method and system based on machine vision | |
| TWI590193B (en) | Image method for classifying insects and insect classifying process | |
| CN106442525B (en) | Online detection method for walnut internal shriveling defect | |
| CN102680414B (en) | Automatic grading device for red globe grapes based on machine vision and method thereof | |
| CN101907453A (en) | Online measurement method and device of dimensions of massive agricultural products based on machine vision | |
| CN104198497A (en) | Surface defect detection method based on visual saliency map and support vector machine | |
| CN108318494B (en) | The red online vision detection and classification devices and methods therefor for proposing fruit powder | |
| CN108020556A (en) | A kind of online damage testing sorting technique of corn seed based on machine vision | |
| CN104198325A (en) | Method for measuring ratio of cut stem to cut tobacco based on computer vision | |
| Wang et al. | Recognition of worm-eaten chestnuts based on machine vision | |
| CN210071686U (en) | Fruit grading plant based on orthogonal binocular machine vision | |
| CN111805541A (en) | Deep learning-based traditional Chinese medicine decoction piece cleaning and selecting device and cleaning and selecting method | |
| CN102788806B (en) | Fruit peel defect detection method based on spheroidic brightness transformation | |
| Thong et al. | Mango classification system uses image processing technology and artificial intelligence | |
| Qiaohua et al. | Grape size detection and online gradation based on machine vision | |
| CN102680488B (en) | Device and method for identifying massive agricultural product on line on basis of PCA (Principal Component Analysis) | |
| Gong et al. | Recent developments of seeds quality inspection and grading based on machine vision | |
| Thong et al. | Mango sorting mechanical system combines image processing | |
| CN111353432A (en) | Rapid honeysuckle medicinal material cleaning method and system based on convolutional neural network | |
| CN202599818U (en) | Machine vision-based red globe grape automatic-classifying device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |